Therapy of aspirin-sensitive asthmatics takes one of two general approaches: desensitization or avoidance. Avoidance of triggering substances seldom alters the clinical course of patients’ asthma. The therapy of asthma has been nonspecific; however, in theory, 5-lipoxygenase inhibitors such as zileuton or leukotriene antagonists should provide specific therapy. A few studies have investigated use of leukotriene modifiers to prevent aspirin-induced bronchospasm in aspirin-sensitive asthmatics.51,52,53,54,55 Pretreatment with zileuton in eight aspirin-sensitive asthma patients protected them from the same threshold-provoking doses of aspirin.51 However, larger, escalating doses of aspirin above the threshold challenge doses were not examined in this study. Furthermore, when doses of aspirin were escalated above the threshold provocative doses, zileuton did not prevent formation of leukotrienes.52 In a similar study, pretreatment with montelukast 10 mg/day did not protect patients when aspirin doses were increased above their threshold doses.46 In another study, the mean provoking dose of aspirin did not differ in the asthmatics who were taking leukotriene modifiers and the control group (60.4 mg vs 70.3 mg, respectively).56 Although initial studies suggested that leukotriene modifiers blocked aspirin-induced reactions, it is now apparent that they merely shift the dose—response curve to the right, leaving the patient at risk at higher doses.57 Thus even patients who might benefit from leukotriene modifiers should avoid aspirin and all NSAIDs. A case of ibuprofen 400-mg-induced asthma was reported in an asthmatic patient on zafirlukast 20 mg twice daily.58 Furthermore, most of the challenge studies are based on incremental doses of aspirin or NSAIDs, and exposure of patients to full clinical doses of aspirin or NSAIDs can overcome the antagonistic effect of leukotriene modifiers. The respiratory symptoms can be decreased but not prevented by pretreatment with antihistamines and cromolyn.59 The long-term asthma control of patients with aspirin sensitivity does not differ from that for other asthmatics. There is no evidence to support that aspirin-sensitive asthmatics respond better to leukotriene modifiers. In a double-blind, randomized, placebo-controlled study, aspirin-sensitive asthmatic patients on montelukast showed a 10% improvement in FEV1 compared with the placebo group.60 Similar results were reported when montelukast was compared with placebo in patients with intermittent or persistent asthma.61
β-Adrenergic receptor blockers comprise the other large class of drugs that can be hazardous to a person with asthma. Even the more cardioselective agents such as acebutolol, atenolol, and metoprolol have been reported to cause asthma attacks.26 Patients with asthma may take nonselective and β1-selective blockers without incident for long periods; however, the occasional report of fatal asthma attacks resistant to therapy with β-agonists should provide ample warning of the dangers inherent in β-blocker therapy.26
If a patient with bronchial hyperreactivity requires β-blocker therapy, one of the selective β1-blockers (eg, acebutolol, atenolol, or metoprolol) should be used at the lowest possible dose. In a meta-analysis of 29 clinical trials in patients with mild to moderate airway obstruction, cardioselective β-blockers did not produce any clinically significant respiratory effects in short-term.62 Similar results were reported in patients with chronic obstructive pulmonary disease (COPD).63 In a large cohort study in United Kingdom, more than 53,000 patients with asthma were identified who were issued oral β-blocker therapy and followed for at least 84 days. The authors did not find a significant difference in asthma exacerbation, defined as use of oral corticosteroid, after prescribing a new oral β-blocker therapy compared to baseline. There was no difference in stratification for β-blocker selectivity in the cohort.64 Celiprolol and betaxolol appear to possess greater cardioselectivity than currently marketed drugs.65,66 Timolol is metabolized by CYP2D6 and patients who are poor metabolizers CYP26 may have higher systemic concentration of timolol resulting in respiratory adverse events.67 Fatal status asthmaticus68 and interstitial lung diseases69 have occurred with the topical administration of the nonselective timolol maleate ophthalmic solution for the treatment of open-angle glaucoma.68 Although ophthalmic betaxolol suggests that it is well tolerated even in timolol-sensitive asthmatics, long-term betaxolol therapy in glaucoma patients with history of pulmonary diseases has been associated with pulmonary obstruction.70,71,72 Airway obstruction following topical β-blockers for glaucoma has also been reported in patients with no history of airway obstruction and close monitoring is warranted.73
Severe, life-threatening asthmatic reactions following consumption of restaurant meals and wine have occurred secondary to ingestion of the food preservative potassium metabisulfite.74,75 Sulfites have been used for centuries as preservatives in wine and food. As antioxidants, they prevent fermentation of wine and discoloration of fruits and vegetables caused by contaminating bacteria.76 Previously, sulfites had been given “generally recognized as safe” status by the Food and Drug Administration (FDA). Sensitive patients react to concentrations ranging from 5 to 100 mg, amounts that are consumed routinely by anyone eating in restaurants. Consumption of sulfites in US diets is estimated to be 2 to 3 mg/day in the home with 5 to 10 mg per 30 mL of beer or wine consumed.75 Anaphylactic or anaphylactoid reactions to sulfites in nonasthmatics are extremely rare. In the general asthmatic population, the overall presence of reactions to sulfites are about 3.9% with more persistent asthma patients at a higher rate.77 Approximately 5% of steroid-dependent asthmatics demonstrate sensitivity to sulfiting agents.76
Three different mechanisms have been proposed to explain the reaction to sulfites in asthmatic patients.76,78 The first is explained by the inhalation of sulfur dioxide, which produces bronchoconstriction in all asthmatics through direct stimulation of afferent parasympathetic irritant receptors. Furthermore, inhalation of atropine or the ingestion of doxepin protects sulfite-sensitive patients from reacting to the ingestion of sulfites. The second theory, IgE-mediated reaction, is supported by reported cases of sulfite-sensitive anaphylaxis reaction in patients with positive sulfite skin test. Finally, a reduced concentration of sulfite oxidase enzyme (the enzyme that catalyzes oxidation of sulfites to sulfates) compared with normal individuals has been demonstrated in a group of sulfite-sensitive asthmatics.
A number of pharmacologic agents contain sulfites as preservatives and antioxidants. The FDA now requires warning labels on drugs containing sulfites. Most manufacturers of drugs for the treatment of asthma have discontinued the use of sulfites. In addition, labeling is required on packaged foods that contain sulfites at 10 parts per million or more, and sulfiting agents are no longer allowed on fresh fruits and vegetables (excluding potatoes) intended for sale.
Pretreatment with cromolyn, anticholinergics, and cyanocobalamin have protected sulfite-sensitive patients.76,79 Presumably, pharmacologic doses of vitamin B12 catalyze the nonenzymatic oxidation of sulfite to sulfate.
Both ethylenediamine tetraacetic acid (EDTA) and benzalkonium chloride, used as stabilizing and bacteriostatic agents, respectively, can produce bronchoconstriction.51 In addition to producing bronchoconstriction, EDTA potentiates the bronchial responsiveness to histamine.80 These effects presumably are mediated through calcium chelation by EDTA. Benzalkonium chloride is more potent than EDTA, and its mechanism appears to be a result of mast cell degranulation and stimulation of irritant C fibers in the airways.80
The bronchoconstriction from benzalkonium chloride can be blocked by cromolyn but not the anticholinergic ipratropium bromide.81 Benzalkonium chloride is found in the commercial multiple-dose nebulizer preparations of ipratropium bromide and beclomethasone dipropionate marketed in the United Kingdom and Europe and is presumed to be in part responsible for paradoxical wheezing following administration of these agents.80,81,82 Benzalkonium chloride is also found in albuterol nebulizer solutions marketed in the United States and has been implicated as a possible cause of paradoxical wheezing in infants receiving this preparation.80 The effect of these agents on FEV1 when used in the amount administered for treatment of acute asthma was evaluated in subjects with stable asthma.83 Patients were assigned randomly to inhale up to four 600-mcg nebulized doses of EDTA and benzalkonium chloride and normal saline. The change in FEV1 was not different between EDTA and the placebo group; however, benzalkonium chloride was associated with a statistically significant decrease in FEV1 compared with placebo. It is important to consider that these agents are always used in combination with bronchodilators and β2-agonists, which are potent mast cell stabilizers, and the anecdotal reports have not yet been confirmed with controlled investigations.70,71,72,73,74,75,76,77,78,79,80,81
Iodinated radiocontrast materials are the most common cause of anaphylactoid reactions producing bronchospasm.84 This is discussed in more detail in Chapter 22.
Natural Rubber Latex Allergy
Allergy to natural rubber latex, first reported in 1989 in the United States, is a common cause of occupational allergy for healthcare workers.85 Natural rubber is a processed plant product from the commercial rubber tree, Hevea brasiliensis.86 Latex allergens are proteins found in both raw latex and the extracts used in finished rubber products. Latex gloves are the largest single source of exposure to the protein allergens.86
The reported prevalence of latex allergy depends on the sample population. In the general population, latex allergy is between 5% and 10%; however, the prevalence increases in healthcare workers to 0.5% to 17%.86,87 Risk factors for latex allergy include frequent exposure to rubber gloves, history of atopic disease, and presence or history of hand dermatitis. Patients with spina bifida are at an increased risk of latex allergy, with an incidence of 18% to 64% as a result of early and repeated exposure to rubber devices during the surgical procedures.86,88,89
Clinical manifestations of latex allergy range from contact dermatitis and urticaria, rhinitis and asthma, and reported cases of anaphylaxis.85,86 The early manifestation of rubber allergy is contact urticaria, which is an IgE-mediated reaction to rubber proteins following direct contact with the medical devices: mainly rubber gloves.86 Contact dermatitis may occur within 1 to 2 days. Contact dermatitis is a cell-mediated delayed-type hypersensitivity reaction to the additive chemical component of rubber products.86 Rhinitis and asthma may follow inhalation of allergens carried by cornstarch powder used to coat the latex gloves. Asthma caused by occupational exposure is seen mostly in atopic patients with histories of seasonal and perennial allergies and asthma.86 Isolated cases of wheezing secondary to latex exposure in patients without a history of asthma have also been reported.86
The diagnosis of latex allergy is based on the presence of latex-specific IgE, as well as symptoms consistent with IgE-mediated reactions.90 The mainstay of therapy for latex allergy is avoidance. Substitution of powdered latex gloves with low protein natural rubber latex has reduced the rate of latex allergy and sensitivity in healthcare workers.91 The FDA requires appropriate labeling for all medical devices containing natural rubber latex to ensure avoidance and a latex-free environment. The role of pretreatment with antihistamines, corticosteroids, and allergen immunotherapy remains to be determined.86,90 Specific immunotherapy for latex allergy (either subcutaneous or sublingual immunotherapy) has been evaluated and sublingual immunotherapy seems more tolerable than the subcutaneous injection; however, systemic reactions have been reported during the build-up phase of immunotherapy92 and it may not be the best option for patients with moderate to severe asthma.87,93
Angiotensin-Converting Enzyme Inhibitor-Induced Cough
Cough has become a well-recognized side effect of angiotensin-converting enzyme (ACE) inhibitor therapy. According to spontaneous reporting by patients, cough occurs in 1% to 10% of patients receiving ACE inhibitors, with a preponderance of females. In a retrospective analysis, 14.6% of women had cough compared with 6.0% of the men on ACE inhibitors. It is suggested that women have a lower cough threshold, resulting in their reporting this adverse effect more commonly than men.94 Studies specifically evaluating cough caused by ACE inhibitors report a prevalence of 19% to 25%.94,95 Patients receiving ACE inhibitors had a 2.3 times greater likelihood of developing cough than a similar group of patients receiving diuretics.94 Patients with hyperreactive airways do not appear to be at greater risk.94,96 African Americans and Chinese have a higher incidence of cough.97 When different disease states were compared, 26% of patients with heart failure had ACE inhibitor-induced cough compared with 14% of those with hypertension.97 Cough can occur with all ACE inhibitors.98
The cough is typically dry and nonproductive, persistent, and not paroxysmal.98 The severity of cough varies from a “tickle” to a debilitating cough with insomnia and vomiting. The cough can begin within 3 days or have a delayed onset of up to 12 months following initiation of ACE inhibitor therapy.98 The cough remits within 1 to 4 days of discontinuing therapy but (rarely) can last up to 4 weeks and recur with rechallenge.98 Patients should be given a 4-day withdrawal to determine if the cough is induced by ACE inhibitors. The chest radiograph is normal, as are pulmonary function tests (spirometry and diffusing capacity). Bronchial hyperreactivity, as measured by histamine and methacholine provocation, may be worsened in patients with underlying bronchial hyperreactivity such as asthma and chronic bronchitis. However, bronchial hyperreactivity is not induced in others.98,99 The cough reflex to capsaicin is enhanced but not to nebulized distilled water or citric acid.98
The mechanism of ACE inhibitor-induced cough is still unknown. ACE is a nonspecific enzyme that also catalyzes the hydrolysis of bradykinin and substance P (see Chapter 13 for more detail) that produce or facilitate inflammation and stimulate lung irritant receptors.98 ACE inhibitors may also induce COX to cause the production of prostaglandins. NSAIDs, benzonatate, inhaled bupivacaine, theophylline, baclofen, thromboxane A2 synthase inhibitor,97,100 and cromolyn sodium all have been used to suppress or inhibit ACE inhibitor-induced cough.98,101 The cough is generally unresponsive to cough suppressants or bronchodilator therapy. No long-term trials evaluating different treatment options for ACE inhibitor-induced cough exist. Cromolyn sodium may be considered first because it is the most studied agent and has minimal toxicity.97 The preferred therapy is withdrawal of the ACE inhibitor and replacement with an alternative antihypertensive agent. Owing to their decrease in ACE inhibitor-induced side effects, angiotensin II receptor antagonists are often recommended in place of an ACE inhibitor; however, there are rare reports of this agent inducing bronchospasm.96,102 The clinical trials suggest that angiotensin II receptor antagonists have the same incidence of cough as placebo. Furthermore, when angiotensin II receptor antagonists were compared with ACE inhibitors, cough occurred much less frequently. Reduction in the incidence of cough with angiotensin II receptor antagonists is likely caused by the lack of effect on clearance of bradykinin and substance P.103 The use of alternative therapies to treat ACE inhibitor-induced cough is generally not recommended.103
Pulmonary edema may result from the failure of any of a number of homeostatic mechanisms. The most common cause of pulmonary edema is an increase in capillary hydrostatic pressure because of left ventricular failure. Excessive fluid administration in compensated and decompensated heart failure patients is the most frequent cause of iatrogenic pulmonary edema. Besides hydrostatic forces, other homeostatic mechanisms that may be disrupted include the osmotic and oncotic pressures in the vasculature, the integrity of the alveolar epithelium, the interstitial pulmonary pressure, and the interstitial lymph flow.8 The edema fluid in cardiogenic pulmonary edema contains a low amount of protein, whereas noncardiogenic pulmonary edema fluid has a high protein concentration.8 This indicates that noncardiogenic pulmonary edema results primarily from disruption of the alveolar epithelium.
The clinical presentation of pulmonary edema includes persistent cough, tachypnea, dyspnea, tachycardia, rales on auscultation, hypoxemia from ventilation—perfusion imbalance and intrapulmonary shunting, widespread fluffy infiltrates on chest roentgenogram, and decreased lung compliance (stiff lungs). Noncardiogenic pulmonary edema may progress to hemorrhage; cellular debris collects in the alveoli, followed by hyperplasia and fibrosis with a residual restrictive mechanical defect.8,104
Narcotic-Induced Pulmonary Edema
The most common drug-induced noncardiogenic pulmonary edema is produced by the narcotic analgesics (Table e30-4).8 Narcotic-induced pulmonary edema is associated most commonly with intravenous heroin use, but also has occurred with morphine, methadone, meperidine, and propoxyphene use.8,104,105 There have also been a few reported cases associated with the use of the opiate antagonist naloxone and nalmefene, a long-acting opioid antagonist.104,106,107 The mechanism is unknown but may be related to hypoxemia similar to the neurogenic pulmonary edema associated with cerebral tumors or trauma or a direct toxic effect on the alveolar capillary membrane.105 Initially thought to occur only with overdoses, most evidence now supports the theory that narcotic-induced pulmonary edema is an idiosyncratic reaction to moderate as well as high narcotic doses.104,105
TABLE e30-4Drugs That Induce Pulmonary Edema
Patients with pulmonary edema may be comatose with depressed respirations or dyspnea and tachypnea. They may or may not have other signs of narcotic overdose. Symptomatology varies from cough and mild crepitations on auscultation with characteristic radiologic findings to severe cyanosis and hypoxemia, even with supplemental oxygen. Symptoms may appear within minutes of intravenous administration but may take up to 2 hours to occur, particularly following oral methadone.105 Hemodynamic studies in the first 24 hours have demonstrated normal pulmonary capillary wedge pressures in the presence of pulmonary edema.
Clinical symptoms generally improve within 24 to 48 hours, and radiologic clearing occurs in 2 to 5 days, but abnormalities in pulmonary function tests may persist for 10 to 12 weeks. Therapy consists of naloxone administration, supplemental oxygen, and ventilatory support if required. Mortality is less than 1%.105
Cough has been reported with intravenous administration of fentanyl in adult and pediatric population.108,109 A cohort of 1,311 adult patients undergoing elective surgery had 120 patients with vigorous cough within 20 seconds after administration of fentanyl. The cough was associated with young age and absence of cigarette smoking.108 Among anesthetic factors, it was associated with the absence of epidurally administered lidocaine and the absence of a priming dose of vecuronium. A history of asthma or COPD had no predictive effect.108 Further clinical trials are required to understand the mechanism of paradoxical cough with fentanyl and to identify the means to prevent it.
Other Drugs That Cause Pulmonary Edema
A paradoxical pulmonary edema has been reported in a few patients following hydrochlorothiazide ingestion but not any other thiazide diuretic.8,110 Acute pulmonary edema rarely has followed the injection of high concentrations of contrast medium into the pulmonary circulation during angiocardiography.8,110 Rare occurrences of pulmonary edema have followed the intravenous administration of bleomycin, cyclophosphamide, and vinblastine.8
The selective β2-adrenergic agonists terbutaline and ritodrine have been reported to induce pulmonary edema when used as tocolytics.8,110 This disorder commonly occurs 48 to 72 hours after tocolytic therapy.107 This has never occurred with their use in asthma patients, even in inadvertent overdosage. This reaction may result from excess fluid administration used to prevent the hypotension from β2-mediated vasodilation or the particular hemodynamics of pregnancy. In a review of 330 patients who received tocolytic therapy and were monitored closely for their fluid status, no episode of pulmonary edema was reported.107
Interleukin-2, a cytokine used alone or in combination with cytotoxic drugs, has been reported to induce pulmonary edema. Although other cytokines have been associated with pulmonary edema, the problem is most significant with interleukin-2. A weight gain of 2 kg has been reported after treatment with interleukin-2.107
Pulmonary edema has occurred occasionally with salicylate overdoses. The serum salicylate concentrations are often greater than 45 mg/dL, and the patients have other signs of toxicity, although some cases have been associated with concentrations in the usual therapeutic range.104,105
Pulmonary infiltrates with eosinophilia (Löffler syndrome) are associated with nitrofurantoin, para-aminosalicylic acid, methotrexate, sulfonamides, tetracycline, chlorpropamide, phenytoin, NSAIDs, and imipramine (Table e30-5).8,110,111,112 The disorder is characterized by fever, nonproductive cough, dyspnea, cyanosis, bilateral pulmonary infiltrates, and eosinophilia in the blood.8 Lung biopsy has revealed perivasculitis with infiltration of eosinophils, macrophages, and proteinaceous edema fluid in the alveoli. The symptoms and eosinophilia generally respond rapidly to withdrawal of the offending drug.
TABLE e30-5Drugs That Induce Pulmonary Infiltrates with Eosinophilia (Löffler Syndrome)
Sulfonamides were first reported as causative agents in users of sulfanilamide vaginal cream.8 para-Aminosalicylic acid frequently produced the syndrome in tuberculosis patients being treated with this agent.8 There are nine reported cases associated with sulfasalazine use in inflammatory bowel disease.111 The drug associated most frequently with this syndrome is nitrofurantoin.8,105 Nitrofurantoin-induced lung disorders appear to be more common in postmenopausal women.105 Lung reactions made up 43% of 921 adverse reactions to nitrofurantoin reported to the Swedish Adverse Drug Reaction Committee between 1966 and 1976.111 No apparent correlation exists between duration of drug exposure and severity or reversibility of the reaction.111 Most cases occur within 1 month of therapy. Typical symptoms include fever, tachypnea, dyspnea, dry cough, and, less commonly, pleuritic chest pain. Radiographic findings include bilateral interstitial infiltrates, predominant in the bases and pleural effusions 25% of the time. Although there are anecdotal reports that steroids are beneficial, the usual rapid improvement following discontinuation of the drugs brings the usefulness of steroids into question. Complete recovery usually occurs within 15 days of withdrawal.
A few cases of pulmonary eosinophilia have been reported in asthmatics treated with cromolyn.8,111 The significance of this is unknown in light of the occasional spontaneous occurrence of pulmonary eosinophilia in asthmatic patients. Cases of acute pneumonitis and eosinophilia have been reported to occur with phenytoin and carbamazepine therapy.111 Patients have had other symptoms of hypersensitivity, including fever and rashes. The symptoms of dyspnea and cough subside following discontinuation of the drug.
Because of the similarity to pulmonary fibrosis, oxygen-induced lung toxicity is reviewed briefly. More extensive reviews on this topic have been published.113,114
The earliest manifestation of oxygen toxicity is substernal pleuritic pain from tracheobronchitis.114 The onset of toxicity follows an asymptomatic period and presents as cough, chest pain, and dyspnea. Early symptoms are usually masked in ventilator-dependent patients. The first noted physiologic change is a decrease in pulmonary compliance caused by reversible atelectasis. Then decreases in vital capacity occur, followed by progressive abnormalities in carbon monoxide diffusing capacity.114 Decreased inspiratory flow rates, reflected in the need for high inspiratory pressures in ventilator-dependent patients, occur as the fractional concentration of inspired oxygen requirement increases. The lungs become progressively stiffer as the ability to oxygenate becomes more compromised.
The fraction of inspired oxygen and duration of exposure are both important determinants of the severity of lung damage. Normal human volunteers can tolerate 100% oxygen at sea level for 24 to 48 hours with minimal to no damage.113 Oxygen concentrations of less than 50% are well tolerated even for extended periods. Inspired oxygen concentrations between 50% and 100% carry a substantial risk of lung damage, and the duration required is inversely proportional to the fraction of inspired oxygen.113 Underlying disease states may alter this relationship. Lung damage may not be lasting and may improve months to years after the exposure.115,116
Oxygen-induced lung damage is generally separated into the acute exudative phase and the subacute or chronic proliferative phase. The acute phase consists of perivascular, peribronchiolar, interstitial, and alveolar edema with alveolar hemorrhage and necrosis of pulmonary endothelium and type I epithelial cells.113 The proliferative phase consists of resorption of the exudates and hyperplasia of interstitial and type II alveolar lining cells. Collagen and elastin deposition in the interstitium of alveolar walls then leads to thickening of the gas-exchange area and the fibrosis.113
The biochemical mechanism of the tissue damage during hyperoxia is the increased production of highly reactive, partially reduced oxygen metabolites (Fig. e30-1).114 These oxidants are normally produced in small quantities during cellular respiration and include the superoxide anion, hydrogen peroxide, the hydroxyl radical, singlet oxygen, and hypochlorous acid.114 Oxygen free radicals are normally formed in phagocytic cells to kill invading microorganisms, but they are also toxic to normal cell components. The oxidants produce toxicity through destructive redox reactions with protein sulfhydryl groups, membrane lipids, and nucleic acids.114
Schematic of the interaction of oxygen radicals and the antioxidant system. (GSH, glutathione; G6PD, glucose-6-phosphate dehydrogenase; NADP, nicotinamide-adenine dinucleotide phosphate; NADPH, reduced NADP.)
The oxidants are products of normal cellular respiration that are normally counterbalanced by an antioxidant defense system that prevents tissue destruction. The antioxidants include superoxide dismutase, catalase, glutathione peroxidase, ceruloplasmin, and α-tocopherol (vitamin E).117 Antioxidants are ubiquitous in the body. Hyperoxia produces toxicity by overwhelming the antioxidant system. There is experimental evidence that a number of drugs and chemicals produce lung toxicity through increasing production of oxidants (eg, bleomycin, cyclophosphamide, nitrofurantoin, and paraquat) and/or by inhibiting the antioxidant system (eg, carmustine, cyclophosphamide, and nitrofurantoin).118,119
A large number of drugs are associated with chronic pulmonary fibrosis with or without a preceding acute pneumonitis (Table e30-6). The cancer chemotherapeutic agents and hematopoietic stem cell transplantation make up the largest group and have been the subject of numerous reviews.118,119,120 Although the mechanisms by which all the drugs produce pneumonitis and fibrosis are not known, the clinical syndrome, pulmonary function abnormalities, and histopathology present a relatively homogeneous pattern.118 The histopathological picture closely resembles oxidant lung damage, and in some experimental cases, oxygen enhances the pulmonary injury.105 Although the terms pulmonary fibrosis or interstitial pneumonitis have been used widely to describe pneumonia after bone marrow transplantation, in 1991, a National Institutes of Health workshop recommended that the term idiopathic pneumonia syndrome (IPS) should be used to avoid histopathological terms and to define the inherent heterogeneity of this disorder.121 IPS accounts for more than 40% of deaths related to bone marrow transplantation.85 Suggested causes of IPS include radiation or chemotherapy regimens prior to transplantation, graft-versus-host disease, unrecognized infections, and other inflammation-related lung injuries.120,122,123 IPS is characterized by dyspnea, hypoxemia, nonproductive cough, diffuse alveolar damage, and interstitial pneumonitis in the absence of lower respiratory infection. IPS has been reported early and late, up to 24 months after bone marrow transplantation.120,123
TABLE e30-6Drugs That Induce Pneumonitis and/or Fibrosis
The lung damage following ingestion of the contact herbicide paraquat classically resembles hyperoxic lung damage. Hyperoxia accelerates the lung damage induced by paraquat. Lung toxicity from paraquat occurs following oral administration in humans and aerosol administration and inhalation in experimental animals.119 The pulmonary specificity of paraquat results in part from its active uptake into lung tissue. Paraquat readily accepts an electron from reduced nicotinamide-adenine dinucleotide phosphate and then is reoxidized rapidly, forming superoxide and other oxygen radicals.119 The toxicity may be a result of nicotinamide-adenine dinucleotide phosphate depletion (see eFig. 30-1) and/or excess oxygen free radical generation with lipid peroxidation. Treatment with exogenous superoxide dismutase has had limited and conflicting results.119
A number of furans have been shown to produce oxidant injury to lungs.119 Occasionally, patients with acute nitrofurantoin lung toxicity will progress to a chronic reaction leading to fibrosis, and rarely, a patient may develop chronic toxicity without an antecedent acute reaction. Like paraquat, nitrofurantoin undergoes cyclic reduction and reoxidation that may produce superoxide radicals or deplete nicotinamide-adenine dinucleotide phosphate. In addition, nitrofurantoin inhibits glutathione reductase, an enzyme involved in the glutathione antioxidant system (see Fig. e30-1). Table e30-7 lists possible nondrug causes of pulmonary fibrosis.
TABLE e30-7Possible Causes of Pulmonary Fibrosis |Favorite Table|Download (.pdf) TABLE e30-7 Possible Causes of Pulmonary Fibrosis
Idiopathic pulmonary fibrosis (fibrosing alveolitis)
Pneumoconiosis (asbestosis, silicosis, coal dust, talc berylliosis)
Hypersensitivity pneumonitis (molds, bacteria, animal proteins, toluene diisocyanate, epoxy resins)
Systemic lupus erythematosus
Byssinosis (cotton workers)
Siderosis (arc welders’ lung)
Chemicals (thioureas, trialkylphosphorothioates, furans)
Drugs (see Tables e30-5, e30-6 and e30-8)
Drugs Associated with Pulmonary Fibrosis
A number of cancer chemotherapeutic agents produce pulmonary fibrosis.124 In an excellent review,118 six predisposing factors for the development of cytotoxic drug-induced pulmonary disease were described: (a) cumulative dose, (b) increased age, (c) concurrent or previous radiotherapy, (d) oxygen therapy, (e) other cytotoxic drug therapy, and (f) preexisting pulmonary disease. Drugs that are directly toxic to the lung would be expected to show a dose–response relationship. Dose–response relationships have been established for bleomycin, busulfan, and carmustine (BCNU).118 Bleomycin and busulfan exhibit threshold cumulative doses below which a very small percentage of patients exhibit toxicity, but carmustine shows a more linear relationship.113 Older patients appear to be more susceptible, possibly as a result of a decrease in the antioxidant defense system. The Childhood Cancer Survivor Study (CCSS), a retrospective cohort of over 14,000 survivors of cancer over 5 years reported a cumulative incidence of pulmonary fibrosis, chronic cough, and shortness of breath with cyclophosphamide, bleomycin, busulfan, BCNU and lomustine (CCNU). The incidence will continue to rise up to 25 years from the time of diagnosis.125 About 1% to 10% of the patients taking bleomycin, carmustine, busulfan, or cyclophosphamide develop lung toxicity.1
Excessive irradiation produces a pneumonitis and fibrosis thought to be caused by oxygen free radical formation.118,120 Evidence for synergistic toxicity with radiation exists for bleomycin, busulfan, and mitomycin. Hyperoxia has shown synergistic toxicity with bleomycin, cyclophosphamide, and mitomycin.118 Carmustine, mitomycin, cyclophosphamide, bleomycin, and methotrexate all appear to show increased lung toxicity when they are part of multiple-drug regimens.
BCNU is associated with the highest incidence of pulmonary toxicity (20%-30%).118 The lung pathology generally resembles that produced by bleomycin and busulfan. Unique to BCNU is the finding of fibrosis in the absence of inflammatory infiltrates. BCNU preferentially inhibits glutathione reductase, the enzyme required to regenerate glutathione, thus reducing glutathione tissue stores.118,119 The patients present with dyspnea, tachypnea, and nonproductive cough that may begin within a month of initiation of therapy but may not develop for as long as 3 years.118 Most patients receiving BCNU develop fibrosis that may remain asymptomatic or become symptomatic any time up to 17 years after therapy.126 The cumulative dose has ranged from 580 to 2,100 mg/m2.119 The disease is usually slowly progressive with a mortality rate from 15% to greater than 90% depending on the study and period of follow-up. In a retrospective study, the risk factors for development of IPS and prognostic factors for outcomes were evaluated in 94 patients with relapsed Hodgkin disease treated with BCNU containing high-dose chemotherapy and hematopoietic support. The risk factors for pulmonary fibrosis and mortality were female sex and dose of BCNU, with all deaths reported in those who received BCNU at doses of more than 475 mg/m2.127 Rapid progression and death within a few days occur in a small percentage of patients.112 Corticosteroids do not appear to be effective in reducing damage.118 Other nitrosoureas, lomustine, and semustine have also been reported to produce lung damage in patients receiving unusually high doses.118
Bleomycin is the best-studied cytotoxic pulmonary toxin. Because of its lack of bone marrow suppression, pulmonary toxicity is the dose-limiting toxicity of bleomycin therapy. The incidence of bleomycin lung toxicity is approximately 4%, which may be affected by the following risk factors: bleomycin cumulative dose, age, high concentration of inspired oxygen, radiation therapy, and multidrug regimens, particularly those with cyclophosphamide.107 Age at the time of treatment with bleomycin may also be a risk factor; patients younger than 7 years at the time of receiving bleomycin therapy are more likely to develop pulmonary toxicity compared with older subjects.107 The cumulative dose above which the incidence of toxicity significantly increases is 450 to 500 units.118 However, rapidly fatal pulmonary toxicity has occurred with doses as low as 100 units.118
Experimentally, bleomycin generates superoxide anions, and the lung toxicity is increased by radiation and hyperoxia.118 Pretreatment with superoxide dismutase and catalase reduces toxicity in experimental animals.118 Bleomycin also oxidizes arachidonic acid, which may account for the marked inflammation. Bleomycin may also affect collagen deposition by its stimulation of fibroblast growth.118 Combination of bleomycin with other cytotoxic agents, particularly regimens containing cyclophosphamide, may predispose patients to pulmonary damage.
There are two distinct clinical patterns of bleomycin pulmonary toxicity. Chronic progressive fibrosis is the most common; acute hypersensitivity reactions occur infrequently. Patients present with cough and dyspnea. The first physiologic abnormality seen is a decreased diffusing capacity of carbon monoxide.118 Chest radiographs show a bibasilar reticular pattern, and gallium scans show marked uptake in the involved lung.118 Chest radiographic changes lag behind pulmonary function abnormalities. Spirometry tests before each bleomycin dose are not predictive of toxicity. The single-breath diffusing capacity of carbon monoxide is the most sensitive indicator of bleomycin-induced lung disease. Although it is not absolutely predictive, a drop of 20% or greater in the diffusing capacity of carbon monoxide is an indication for using alternative therapies.118 The prognosis of bleomycin lung toxicity has improved as a consequence of early detection, but the mortality rate is approximately 25%. Mild cases respond to discontinuation of bleomycin therapy.107 Corticosteroid therapy appears to be helpful in patients with acute pneumonitis, although there have been no controlled trials. Patients with chronic fibrosis are less likely to respond. Although corticosteroids have been used for a number of drug-induced pulmonary problems, a study in mice showing a potential for worsening of lung damage when administered early during the repair stage should sound a word of caution against their indiscriminate use.128 Current clinical trials do not support use of glucocorticoids in prevention, early, or late phases of acute lung injury or acute respiratory distress.129
Mitomycin is an alkylating antibiotic that produces pulmonary fibrosis at a frequency of 3% to 12%.118 The mechanism is unknown, but oxygen and radiation therapy appear to enhance the development of toxicity.118 The clinical presentation and symptoms are the same as for bleomycin. The mortality rate is approximately 50%. Early withdrawal of the drug and administration of corticosteroids appear to improve the outcome significantly. In a prospective trial, routine pulmonary function test monitoring did not appear to be predictive of pulmonary toxicity.130
A number of alkylating agents are associated with pulmonary fibrosis (see Table 30-5). The incidence of clinical toxicity is around 4%, although subclinical damage is apparent in up to 46% of patients at autopsy. The mechanism of toxicity is unknown; however, epithelial cell damage that triggers the arachidonic acid inflammatory cascade may be the initiating event.118 The clinical presentation is insidious, with 4 years being the average duration of therapy before the onset of symptoms. Patients present with low-grade fever, weight loss, weakness, dyspnea, cough, and rales.118 Pulmonary function tests initially show abnormal diffusion capacity followed by a restrictive pattern (low vital capacity). The histopathologic findings are nonspecific. The prognosis is one of slow progression with a mean survival of 5 months following diagnosis.118 Although there is no direct dose-dependent correlation, patients receiving less than 500 mg of busulfan do not develop the syndrome without concomitant radiation or use of other pulmonary toxic chemotherapeutic agents.118 There are anecdotal reports of beneficial responses to corticosteroids, but no controlled studies have been done.
Cyclophosphamide infrequently produces pulmonary toxicity.1 More than 20 well-documented cases have been reported to date. In animal models, cyclophosphamide produces reactive oxygen radicals. High oxygen concentrations produce synergistic toxicity with cyclophosphamide. The duration of therapy before the onset of symptoms is highly variable, and there may be a delay of several months between the onset of symptoms and discontinuation of the drug.118 Cyclophosphamide may potentiate carmustine lung toxicity.118 Clinical symptoms usually consist of dyspnea on exertion, cough, and fever. Inspiratory crackles and the bibasilar reticular pattern typical of cytotoxic drug-induced radiographic changes are present. Histopathological changes are also nonspecific. Approximately 60% of patients recover. Corticosteroid therapy has been reported to be beneficial; however, death despite corticosteroid administration has also been reported.
Chlorambucil, melphalan, and uracil mustard are also associated with pulmonary fibrosis. Of the alkylating agents, only nitrogen mustard and thiotepa have not been reported to cause fibrotic pulmonary toxicity.118
Methotrexate was first reported to induce pulmonary toxicity in 1969.118 The pulmonary toxicity to methotrexate is unique in that discontinuation is not always necessary, and reinstitution of the drug may not produce recurrence of symptoms.8 Methotrexate pulmonary toxicity most commonly appears to result from hypersensitivity,1, 111 and it can occur 3 or more years following methotrexate therapy.131 Age, sex, underlying pulmonary disease, duration of therapy, or smoking is not associated with an increased risk of pneumonitis with methotrexate.131 Serial pulmonary function tests did not help identify pneumonitis in patients receiving methotrexate before the onset of clinical symptoms.131 Reductions in diffusing capacity of carbon monoxide and lung volumes are the most common manifestations of methotrexate lung toxicity.107 Pulmonary edema and eosinophilia are common, and fibrosis occurs in only 10% of the patients who develop acute pneumonitis.118 Systemic symptoms of chills, fever, and malaise are common before the onset of dyspnea, cough, and acute pleuritic chest pain. Methotrexate is also associated with granuloma formation.118
The prognosis of methotrexate-induced pulmonary toxicity is good, with a 1% or less mortality rate.111 Pulmonary toxicity has followed intrathecal as well as oral administration and has occurred after single doses as well as long-term daily and intermittent administration. Pneumonitis has been reported to occur up to 4 weeks following discontinuation of therapy.118 Numerous anecdotal reports have claimed dramatic benefit from corticosteroid therapy. It is unknown whether intermittent (weekly) dosing, as is done for rheumatoid arthritis, decreases the risk of methotrexate-induced pulmonary toxicity because pneumonitis has occurred with this form of dosing.
Rarely, azathioprine and its major metabolite 6-mercaptopurine have been reported to produce an acute restrictive lung disease. Procarbazine, a methylhydrazine associated more commonly with Löffler syndrome, rarely has been associated with pulmonary fibrosis.111 The vinca alkaloids vinblastine and vindesine have been reported to produce severe respiratory toxicity in association with mitomycin. The incidence with the combination is 39% and may represent a true synergistic effect between these agents.118 The safety profile of gemcitabine was reviewed in 22 completed clinical trials with more than 900 patients and pulmonary toxicity was rare at a rate of 1.4%.132 Gemcitabine has been reported to cause noncardiogenic pulmonary edema and use of corticosteroids and diuretics should be considered early on to prevent mortality.133
Pulmonary fibrosis associated with the ganglionic-blocking agent hexamethonium was first reported in 1954 (see Table e30-6).8 Patients developed extreme dyspnea after several months on the drug. Pathological findings were consistent with bronchiectasis, bronchiolectasis, and fibrosis.8 This phenomenon has occurred occasionally with use of the other ganglionic blockers (ie, mecamylamine and pentolinium).8
In 1959, radiographic changes characteristic of diffuse pulmonary fibrosis were reported in 27 (87%) of 31 patients who had taken phenytoin for 2 years or more.105 Since then, studies have been conflicting. If phenytoin does produce chronic fibrosis, it would appear to be a relatively rare event.
Gold salts (sodium aurothiomalate) used in the treatment of rheumatoid arthritis have produced pulmonary fibrosis with cough, dyspnea, and pleuritic pain 5 to 16 weeks following institution of therapy.105 Pulmonary function tests show a restrictive defect, and patients generally have an eosinophilia. The reactions improve on discontinuation of the gold therapy and recur promptly on reexposure. The pulmonary deficit may not resolve completely.
With recent increase in approval of biological including monoclonal antibody agents, the reports of interstitial lung diseases (with no identified pattern) are increasing; some of these agents include tumor necrosis factor-α class of medication, recombinant interferons, rituximab, cetuximab, bevacizumab, alemtuzumab or traztuzumab.134,135,136
Amiodarone, a benzofuran derivative, produces pulmonary fibrosis when used for supraventricular and ventricular arrhythmias (see Table e30-6).137 The duration of amiodarone therapy before the onset of symptoms has ranged from 4 weeks to 6 years.105,137,138 The estimated incidence is 1 in 1,000 to 2,000 treated patients per year. Approximately 6% of the patients taking amiodarone will have pulmonary abnormalities with mortality rate of 10% to 20%.1 The clinical course is variable, ranging from acute onset of dyspnea with rapid progression into severe respiratory failure and death caused by slowly developing exertional dyspnea over a few months. Patients generally improve on discontinuation of the drug.137,138 The majority of patients develop reactions while taking maintenance doses greater than 400 mg daily for more than 2 months or smaller doses for more than 2 years. The risk of amiodarone pulmonary toxicity is higher during the first 12 months of therapy even at a low dosage.139 Other risk factors include cardiopulmonary surgery combined with the administration of high concentrations of oxygen,139 maintenance dose, cumulative dose of amiodarone, and age.140 The prevalence of lung toxicity increases from 4.2% to 10.6% from the first to the fifth year of amiodarone use. Patients 60 years or older have a threefold increase in risk of toxicity for each subsequent decade compared to those younger than 60 years.140 Pulmonary function including DLCO at baseline and routinely or for unexpected pulmonary symptoms is recommended. A reduction in DLCO of 15% has a sensitivity of 68% to 100% and a specificity of 69% to 95% to diagnose pulmonary toxicity.141,142 Clinical findings include exertional dyspnea, nonproductive cough, weight loss, and occasionally low-grade fever.105,138 Radiographic changes are nondiagnostic and consist of diffuse bilateral interstitial changes consistent with a pneumonitis. Pulmonary function abnormalities include hypoxia, restrictive changes, and diffusion abnormalities.
The mechanism of amiodarone-induced pulmonary toxicity is multifactorial. Amiodarone and its metabolite can damage lung tissue directly by a cytotoxic process or indirectly by immunologic reactions.139,140 Amiodarone is an amphiphilic molecule that contains both a highly apolar aromatic ring system and a polar side chain with a positively charged nitrogen atom.137 Amphiphilic drugs characteristically produce a phospholipid storage disorder in the lungs of experimental animals and humans.119 Chlorphentermine, an anorectic, is the prototype amphiphilic compound. The mechanism is currently believed to be the inhibition of lysosomal phospholipases.119 The inflammation and fibrosis are thought to be a late finding resulting from nonspecific inflammation following the breakdown of phospholipid-laden macrophages.137
In a review of 39 cases, 9 patients died, and the remaining 30 patients had resolution of abnormalities after withdrawal of the drug.137 Some patients have had resolution with lowering of the dosage, and therapy has been reinstituted at lower doses without problems in others. Of the patients who died, one half had received corticosteroids. There are reports of a protective effect with prophylactic corticosteroids and other reports of patients developing amiodarone lung toxicity while on corticosteroids.137 The use of corticosteroids for months to 1 year after stopping amiodarone is recommended, despite the lack of controlled trials.143
Miscellaneous Pulmonary Toxicity
Drugs may produce serious pulmonary toxicity as part of a more generalized disorder. The pleural thickening, effusions, and fibrosis that occur as an extension of the retroperitoneal fibrotic reactions of methysergide and practolol or as part of a drug-induced lupus syndrome are the most common examples (Table e30-8).
TABLE e30-8Drugs That May Induce Pleural Effusions and Fibrosis
Pleural and pulmonary fibrosis has been reported in one patient taking pindolol, a β-blocker structurally similar to practolol, an agent known to produce fibrosis.73 Acute pleuritis with pleural effusions and fibrosis is a prominent manifestation of drug-induced lupus syndrome. Procainamide is associated with the largest number of pulmonary reactions, with 46% of patients with the lupus syndrome developing pulmonary complications.8 Symptoms include pleuritic pain and fever with muscle and joint pain. Chest radiographs show bilateral pleural effusions and linear atelectasis. Patients have a positive antinuclear antibody test. Symptoms usually resolve within 6 weeks of drug withdrawal.8
Hydralazine is the next most common cause of lupus syndrome. Most patients who develop pleuropulmonary manifestations have antecedent symptoms of generalized lupus.8 Other drugs that produce the lupus syndrome include isoniazid and phenytoin. Phenytoin can also produce hilar lymphadenopathy as part of a generalized pseudolymphoma or lymphadenopathy syndrome.8
Monitoring Therapeutic Outcomes
Monitoring for drug-induced pulmonary diseases consists primarily of having a high index of suspicion that a particular syndrome may be drug induced. Presently, there is no defined diagnostic workup in patient with suspected drug-induced pulmonary disease. Most hypersensitivity or allergic reactions (bronchospasm) occur rapidly, within the first 2 weeks of therapy with the offending agent, and reverse rapidly with appropriate therapy (eg, withdrawal of the offending agent and administration of corticosteroids and bronchodilators). Dyspnea associated with Löffler syndrome and acute pulmonary edema syndromes also improve rapidly in 1 to 2 days. However, some residual defect in diffusion capacity and the roentgenogram may persist for a few weeks. It is probably unnecessary to do follow-up spirometry or diffusion capacity determinations in these patients unless there is some concern that the syndrome will progress to pulmonary fibrosis (through the use of bleomycin or nitrofurantoin).1
The routine monitoring of patients receiving known pulmonary toxins with dose-dependent toxicity such as amiodarone, bleomycin, or carmustine is still controversial. For chronic fibrosis, the diffusing capacity of carbon monoxide is the most sensitive test and may be useful in patients receiving bleomycin for detecting and preventing further deterioration of lung function with continued administration. Carmustine lung toxicity may be delayed up to 10 years following administration, and routine monitoring has not proved preventive. Monitoring patients receiving amiodarone in doses greater than 400 mg/day every 4 to 6 months may prove useful in detecting early disease that requires lowering the amiodarone or stopping the drug. Because there is no evidence of a cumulative dose effect once it has been established that the patient can tolerate the elevated dose, continued routine monitoring past the first year is unnecessary.